Re: Canon's support and information.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



there are 2 types of dead pixels... black and white... black are those that does not respond to any light... these are very easy to find during the factories' test and calibration phase... white pixels on the other hand are harder to detect as they are mixed with the bias noise and a whole bunch of other stuff...


Also following Alfred Tay's post and your rational, since dead pixels are
white (Alfred refers to them as WBP or White Bad Pixels) do you think it
would be possible to do the following to test the number of bad pixels in
the sensor:

1. Photograph a subject of an even color with RGB values that contrast
reasonably from the RGB values for white (usually closer or around 255).

no need for this... just put the lens cap on and set the exposeure to 2" at various ISO... you can see the noise and the WBP for each level of GAIN (ISO increase are simply base gain levels of the CCE output... in way... ISO does not really make sense... what is more important is the gain level... which is another monster althogether... we shall stick to ISO here)

2. Open the image in some type of text editor or perhaps the EXIF software
you mention to do a search for number values that fall outside of the
average value for the even-colored subject that was photographed.

3. Identify the number of pixels with values that do not coincide with the
values relative to the background. One of Kodak's 8"x10" could be used as
the neutral gray has a known value and would make it simpler to identify the
outstanding pixels.

once you have shot one black frame... you need to repeat this about 10-20 times depending on how detailed u want the final WBP chart to be...WHY? cause u want to differentiate the noise from the system from the real WBP... system noise and bias are random so with enough repetition the random noise will be eliminated leaving u with the FIXED WBP... overlay everything together in PS and you will get a map of all the remaining WBP that the factory has failed in calibration (WHICH is very frequently the case even though each camera is supposed to undergo a WBP calibration prior to release).

BUT the question is SO what? unless we have access to the FW calibration which we dun... we cannot do anything except to complain to the TECH support... even with this we cannot make any compensation of any sorts as these are done to the raw image before demosaic and any image processing...

what raw we are seeing with any of those RAW software is not raw... it all has been demosaic'ed we need access to each individual colour channel before intrepolation... so u see each pixel as it is on the CCD... but again without access to FW we cannot do anything... even SINAR's raw software does nto give u this access even though it is one of the most tweakable raw system...

for a P&S about 5mp... at 100 iso the acceptable number of WBP is about 10,000... at 200 iso about 20,000... the number increases with the gain level... each camera has its WBP map calibrated and stored in the camera... so if u can crack it u will be able to see how messed up it all is...

surely u dun expect me to count each bad pixel??? hahahah no of course not... this is where MATLAB comes in... if you need help with a count... let me know...

so what is the POINT? the point is that once u do this you will know how bad the WBP is on your particular CCD... if it exceeds 10,000 on your CCD then it is time to send your camera back for recalibration... tell them you need a recalibration of your noise & WBP... bugged them till u get it... if one service personnel says no... call until u get someone who says yes... or keep complaing till u get through to management... camera companies like to keep defects quiet... look at the current CCD recall... SONY knew about it for at least two years now... frankly they could have kept quiet and no one would have been the wiser as most of the CCD error happens after the warranty ran dry... must have been some politiking going on up there in the clouds... anyways...thats about all...

i am in the mist of resolving this same issue... if i have more info... i will post it on...

peace
alfred

_________________________________________________________________
Take a break! Find destinations on MSN Travel. http://www.msn.com.sg/travel/


[Index of Archives] [Share Photos] [Epson Inkjet] [Scanner List] [Gimp Users] [Gimp for Windows]

  Powered by Linux